Intrusion detection systems (IDS) play a crucial role in safeguarding modern network infrastructures. The growing sophistication of cyberattacks, coupled with the increasing volume an‎d velocity of network traffic, has led to a rising deman‎d for effective real-time detection capabilities. Deep learning models have demonstrated significant promise in this domain by learning complex patterns in large-scale network data an‎d improving detection accuracy. This literature review provides a comprehensive examination of deep learning models applied to real-time intrusion detection, analyzing their architectures, datasets, eva‎luation metrics, an‎d real-world deployment challenges. We categorize recent approaches based on the deep learning techniques used, including convolutional neural networks (CNNs), recurrent neural networks (RNNs), long short-term memory networks (LSTMs), autoencoders, an‎d hybrid models. Furthermore, we highlight performance comparisons, discuss datasets such as NSL-KDD, CICIDS2017, an‎d UNSW-NB15, an‎d outline open research challenges including explainability, adversarial robustness, an‎d scalability. Finally, this review proposes future research directions aimed at enhancing the practicality an‎d reliability of real-time intrusion detection systems using deep learning. (Ullah et al., 2022; Zhao & Huang, 2024) (Rahman & Chen, 2023; Al-Dhief et al., 2024)

10/29/2025 12:00:00 AM

88069

1404/08/07

Real-Time Intrusion Detection Using Deep Learning Models: A Literature Review

تشخيص نفوذ بلادرنگ , مدل‌هاي يادگيري عميق , امنيت سايبري , امنيت شبكه , IDS (سيستم‌هاي تشخيص نفوذ) , CNN (شبكه‌هاي عصبي كانولوشن)

Real-Time Intrusion Detection , Deep Learning Models , Cybersecurity , Network Security , IDS (Intrusion Detection Systems) , CNN (Convolutional Neural Networks)